sched: clean up wakeup balancing, move wake_affine()
[wrt350n-kernel.git] / fs / ocfs2 / suballoc.c
blob72c198a004df46744c1c025e253763727ce635bb
1 /* -*- mode: c; c-basic-offset: 8; -*-
2 * vim: noexpandtab sw=8 ts=8 sts=0:
4 * suballoc.c
6 * metadata alloc and free
7 * Inspired by ext3 block groups.
9 * Copyright (C) 2002, 2004 Oracle. All rights reserved.
11 * This program is free software; you can redistribute it and/or
12 * modify it under the terms of the GNU General Public
13 * License as published by the Free Software Foundation; either
14 * version 2 of the License, or (at your option) any later version.
16 * This program is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19 * General Public License for more details.
21 * You should have received a copy of the GNU General Public
22 * License along with this program; if not, write to the
23 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
24 * Boston, MA 021110-1307, USA.
27 #include <linux/fs.h>
28 #include <linux/types.h>
29 #include <linux/slab.h>
30 #include <linux/highmem.h>
32 #define MLOG_MASK_PREFIX ML_DISK_ALLOC
33 #include <cluster/masklog.h>
35 #include "ocfs2.h"
37 #include "alloc.h"
38 #include "dlmglue.h"
39 #include "inode.h"
40 #include "journal.h"
41 #include "localalloc.h"
42 #include "suballoc.h"
43 #include "super.h"
44 #include "sysfile.h"
45 #include "uptodate.h"
47 #include "buffer_head_io.h"
49 static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg);
50 static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe);
51 static inline u16 ocfs2_find_victim_chain(struct ocfs2_chain_list *cl);
52 static int ocfs2_block_group_fill(handle_t *handle,
53 struct inode *alloc_inode,
54 struct buffer_head *bg_bh,
55 u64 group_blkno,
56 u16 my_chain,
57 struct ocfs2_chain_list *cl);
58 static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
59 struct inode *alloc_inode,
60 struct buffer_head *bh);
62 static int ocfs2_cluster_group_search(struct inode *inode,
63 struct buffer_head *group_bh,
64 u32 bits_wanted, u32 min_bits,
65 u16 *bit_off, u16 *bits_found);
66 static int ocfs2_block_group_search(struct inode *inode,
67 struct buffer_head *group_bh,
68 u32 bits_wanted, u32 min_bits,
69 u16 *bit_off, u16 *bits_found);
70 static int ocfs2_claim_suballoc_bits(struct ocfs2_super *osb,
71 struct ocfs2_alloc_context *ac,
72 handle_t *handle,
73 u32 bits_wanted,
74 u32 min_bits,
75 u16 *bit_off,
76 unsigned int *num_bits,
77 u64 *bg_blkno);
78 static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh,
79 int nr);
80 static inline int ocfs2_block_group_set_bits(handle_t *handle,
81 struct inode *alloc_inode,
82 struct ocfs2_group_desc *bg,
83 struct buffer_head *group_bh,
84 unsigned int bit_off,
85 unsigned int num_bits);
86 static inline int ocfs2_block_group_clear_bits(handle_t *handle,
87 struct inode *alloc_inode,
88 struct ocfs2_group_desc *bg,
89 struct buffer_head *group_bh,
90 unsigned int bit_off,
91 unsigned int num_bits);
93 static int ocfs2_relink_block_group(handle_t *handle,
94 struct inode *alloc_inode,
95 struct buffer_head *fe_bh,
96 struct buffer_head *bg_bh,
97 struct buffer_head *prev_bg_bh,
98 u16 chain);
99 static inline int ocfs2_block_group_reasonably_empty(struct ocfs2_group_desc *bg,
100 u32 wanted);
101 static inline u32 ocfs2_desc_bitmap_to_cluster_off(struct inode *inode,
102 u64 bg_blkno,
103 u16 bg_bit_off);
104 static inline void ocfs2_block_to_cluster_group(struct inode *inode,
105 u64 data_blkno,
106 u64 *bg_blkno,
107 u16 *bg_bit_off);
109 void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac)
111 struct inode *inode = ac->ac_inode;
113 if (inode) {
114 if (ac->ac_which != OCFS2_AC_USE_LOCAL)
115 ocfs2_inode_unlock(inode, 1);
117 mutex_unlock(&inode->i_mutex);
119 iput(inode);
121 if (ac->ac_bh)
122 brelse(ac->ac_bh);
123 kfree(ac);
126 static u32 ocfs2_bits_per_group(struct ocfs2_chain_list *cl)
128 return (u32)le16_to_cpu(cl->cl_cpg) * (u32)le16_to_cpu(cl->cl_bpc);
131 /* somewhat more expensive than our other checks, so use sparingly. */
132 int ocfs2_check_group_descriptor(struct super_block *sb,
133 struct ocfs2_dinode *di,
134 struct ocfs2_group_desc *gd)
136 unsigned int max_bits;
138 if (!OCFS2_IS_VALID_GROUP_DESC(gd)) {
139 OCFS2_RO_ON_INVALID_GROUP_DESC(sb, gd);
140 return -EIO;
143 if (di->i_blkno != gd->bg_parent_dinode) {
144 ocfs2_error(sb, "Group descriptor # %llu has bad parent "
145 "pointer (%llu, expected %llu)",
146 (unsigned long long)le64_to_cpu(gd->bg_blkno),
147 (unsigned long long)le64_to_cpu(gd->bg_parent_dinode),
148 (unsigned long long)le64_to_cpu(di->i_blkno));
149 return -EIO;
152 max_bits = le16_to_cpu(di->id2.i_chain.cl_cpg) * le16_to_cpu(di->id2.i_chain.cl_bpc);
153 if (le16_to_cpu(gd->bg_bits) > max_bits) {
154 ocfs2_error(sb, "Group descriptor # %llu has bit count of %u",
155 (unsigned long long)le64_to_cpu(gd->bg_blkno),
156 le16_to_cpu(gd->bg_bits));
157 return -EIO;
160 if (le16_to_cpu(gd->bg_chain) >=
161 le16_to_cpu(di->id2.i_chain.cl_next_free_rec)) {
162 ocfs2_error(sb, "Group descriptor # %llu has bad chain %u",
163 (unsigned long long)le64_to_cpu(gd->bg_blkno),
164 le16_to_cpu(gd->bg_chain));
165 return -EIO;
168 if (le16_to_cpu(gd->bg_free_bits_count) > le16_to_cpu(gd->bg_bits)) {
169 ocfs2_error(sb, "Group descriptor # %llu has bit count %u but "
170 "claims that %u are free",
171 (unsigned long long)le64_to_cpu(gd->bg_blkno),
172 le16_to_cpu(gd->bg_bits),
173 le16_to_cpu(gd->bg_free_bits_count));
174 return -EIO;
177 if (le16_to_cpu(gd->bg_bits) > (8 * le16_to_cpu(gd->bg_size))) {
178 ocfs2_error(sb, "Group descriptor # %llu has bit count %u but "
179 "max bitmap bits of %u",
180 (unsigned long long)le64_to_cpu(gd->bg_blkno),
181 le16_to_cpu(gd->bg_bits),
182 8 * le16_to_cpu(gd->bg_size));
183 return -EIO;
186 return 0;
189 static int ocfs2_block_group_fill(handle_t *handle,
190 struct inode *alloc_inode,
191 struct buffer_head *bg_bh,
192 u64 group_blkno,
193 u16 my_chain,
194 struct ocfs2_chain_list *cl)
196 int status = 0;
197 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
198 struct super_block * sb = alloc_inode->i_sb;
200 mlog_entry_void();
202 if (((unsigned long long) bg_bh->b_blocknr) != group_blkno) {
203 ocfs2_error(alloc_inode->i_sb, "group block (%llu) != "
204 "b_blocknr (%llu)",
205 (unsigned long long)group_blkno,
206 (unsigned long long) bg_bh->b_blocknr);
207 status = -EIO;
208 goto bail;
211 status = ocfs2_journal_access(handle,
212 alloc_inode,
213 bg_bh,
214 OCFS2_JOURNAL_ACCESS_CREATE);
215 if (status < 0) {
216 mlog_errno(status);
217 goto bail;
220 memset(bg, 0, sb->s_blocksize);
221 strcpy(bg->bg_signature, OCFS2_GROUP_DESC_SIGNATURE);
222 bg->bg_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
223 bg->bg_size = cpu_to_le16(ocfs2_group_bitmap_size(sb));
224 bg->bg_bits = cpu_to_le16(ocfs2_bits_per_group(cl));
225 bg->bg_chain = cpu_to_le16(my_chain);
226 bg->bg_next_group = cl->cl_recs[my_chain].c_blkno;
227 bg->bg_parent_dinode = cpu_to_le64(OCFS2_I(alloc_inode)->ip_blkno);
228 bg->bg_blkno = cpu_to_le64(group_blkno);
229 /* set the 1st bit in the bitmap to account for the descriptor block */
230 ocfs2_set_bit(0, (unsigned long *)bg->bg_bitmap);
231 bg->bg_free_bits_count = cpu_to_le16(le16_to_cpu(bg->bg_bits) - 1);
233 status = ocfs2_journal_dirty(handle, bg_bh);
234 if (status < 0)
235 mlog_errno(status);
237 /* There is no need to zero out or otherwise initialize the
238 * other blocks in a group - All valid FS metadata in a block
239 * group stores the superblock fs_generation value at
240 * allocation time. */
242 bail:
243 mlog_exit(status);
244 return status;
247 static inline u16 ocfs2_find_smallest_chain(struct ocfs2_chain_list *cl)
249 u16 curr, best;
251 best = curr = 0;
252 while (curr < le16_to_cpu(cl->cl_count)) {
253 if (le32_to_cpu(cl->cl_recs[best].c_total) >
254 le32_to_cpu(cl->cl_recs[curr].c_total))
255 best = curr;
256 curr++;
258 return best;
262 * We expect the block group allocator to already be locked.
264 static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
265 struct inode *alloc_inode,
266 struct buffer_head *bh)
268 int status, credits;
269 struct ocfs2_dinode *fe = (struct ocfs2_dinode *) bh->b_data;
270 struct ocfs2_chain_list *cl;
271 struct ocfs2_alloc_context *ac = NULL;
272 handle_t *handle = NULL;
273 u32 bit_off, num_bits;
274 u16 alloc_rec;
275 u64 bg_blkno;
276 struct buffer_head *bg_bh = NULL;
277 struct ocfs2_group_desc *bg;
279 BUG_ON(ocfs2_is_cluster_bitmap(alloc_inode));
281 mlog_entry_void();
283 cl = &fe->id2.i_chain;
284 status = ocfs2_reserve_clusters(osb,
285 le16_to_cpu(cl->cl_cpg),
286 &ac);
287 if (status < 0) {
288 if (status != -ENOSPC)
289 mlog_errno(status);
290 goto bail;
293 credits = ocfs2_calc_group_alloc_credits(osb->sb,
294 le16_to_cpu(cl->cl_cpg));
295 handle = ocfs2_start_trans(osb, credits);
296 if (IS_ERR(handle)) {
297 status = PTR_ERR(handle);
298 handle = NULL;
299 mlog_errno(status);
300 goto bail;
303 status = ocfs2_claim_clusters(osb,
304 handle,
306 le16_to_cpu(cl->cl_cpg),
307 &bit_off,
308 &num_bits);
309 if (status < 0) {
310 if (status != -ENOSPC)
311 mlog_errno(status);
312 goto bail;
315 alloc_rec = ocfs2_find_smallest_chain(cl);
317 /* setup the group */
318 bg_blkno = ocfs2_clusters_to_blocks(osb->sb, bit_off);
319 mlog(0, "new descriptor, record %u, at block %llu\n",
320 alloc_rec, (unsigned long long)bg_blkno);
322 bg_bh = sb_getblk(osb->sb, bg_blkno);
323 if (!bg_bh) {
324 status = -EIO;
325 mlog_errno(status);
326 goto bail;
328 ocfs2_set_new_buffer_uptodate(alloc_inode, bg_bh);
330 status = ocfs2_block_group_fill(handle,
331 alloc_inode,
332 bg_bh,
333 bg_blkno,
334 alloc_rec,
335 cl);
336 if (status < 0) {
337 mlog_errno(status);
338 goto bail;
341 bg = (struct ocfs2_group_desc *) bg_bh->b_data;
343 status = ocfs2_journal_access(handle, alloc_inode,
344 bh, OCFS2_JOURNAL_ACCESS_WRITE);
345 if (status < 0) {
346 mlog_errno(status);
347 goto bail;
350 le32_add_cpu(&cl->cl_recs[alloc_rec].c_free,
351 le16_to_cpu(bg->bg_free_bits_count));
352 le32_add_cpu(&cl->cl_recs[alloc_rec].c_total, le16_to_cpu(bg->bg_bits));
353 cl->cl_recs[alloc_rec].c_blkno = cpu_to_le64(bg_blkno);
354 if (le16_to_cpu(cl->cl_next_free_rec) < le16_to_cpu(cl->cl_count))
355 le16_add_cpu(&cl->cl_next_free_rec, 1);
357 le32_add_cpu(&fe->id1.bitmap1.i_used, le16_to_cpu(bg->bg_bits) -
358 le16_to_cpu(bg->bg_free_bits_count));
359 le32_add_cpu(&fe->id1.bitmap1.i_total, le16_to_cpu(bg->bg_bits));
360 le32_add_cpu(&fe->i_clusters, le16_to_cpu(cl->cl_cpg));
362 status = ocfs2_journal_dirty(handle, bh);
363 if (status < 0) {
364 mlog_errno(status);
365 goto bail;
368 spin_lock(&OCFS2_I(alloc_inode)->ip_lock);
369 OCFS2_I(alloc_inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
370 fe->i_size = cpu_to_le64(ocfs2_clusters_to_bytes(alloc_inode->i_sb,
371 le32_to_cpu(fe->i_clusters)));
372 spin_unlock(&OCFS2_I(alloc_inode)->ip_lock);
373 i_size_write(alloc_inode, le64_to_cpu(fe->i_size));
374 alloc_inode->i_blocks = ocfs2_inode_sector_count(alloc_inode);
376 status = 0;
377 bail:
378 if (handle)
379 ocfs2_commit_trans(osb, handle);
381 if (ac)
382 ocfs2_free_alloc_context(ac);
384 if (bg_bh)
385 brelse(bg_bh);
387 mlog_exit(status);
388 return status;
391 static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
392 struct ocfs2_alloc_context *ac,
393 int type,
394 u32 slot)
396 int status;
397 u32 bits_wanted = ac->ac_bits_wanted;
398 struct inode *alloc_inode;
399 struct buffer_head *bh = NULL;
400 struct ocfs2_dinode *fe;
401 u32 free_bits;
403 mlog_entry_void();
405 alloc_inode = ocfs2_get_system_file_inode(osb, type, slot);
406 if (!alloc_inode) {
407 mlog_errno(-EINVAL);
408 return -EINVAL;
411 mutex_lock(&alloc_inode->i_mutex);
413 status = ocfs2_inode_lock(alloc_inode, &bh, 1);
414 if (status < 0) {
415 mutex_unlock(&alloc_inode->i_mutex);
416 iput(alloc_inode);
418 mlog_errno(status);
419 return status;
422 ac->ac_inode = alloc_inode;
424 fe = (struct ocfs2_dinode *) bh->b_data;
425 if (!OCFS2_IS_VALID_DINODE(fe)) {
426 OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
427 status = -EIO;
428 goto bail;
430 if (!(fe->i_flags & cpu_to_le32(OCFS2_CHAIN_FL))) {
431 ocfs2_error(alloc_inode->i_sb, "Invalid chain allocator %llu",
432 (unsigned long long)le64_to_cpu(fe->i_blkno));
433 status = -EIO;
434 goto bail;
437 free_bits = le32_to_cpu(fe->id1.bitmap1.i_total) -
438 le32_to_cpu(fe->id1.bitmap1.i_used);
440 if (bits_wanted > free_bits) {
441 /* cluster bitmap never grows */
442 if (ocfs2_is_cluster_bitmap(alloc_inode)) {
443 mlog(0, "Disk Full: wanted=%u, free_bits=%u\n",
444 bits_wanted, free_bits);
445 status = -ENOSPC;
446 goto bail;
449 status = ocfs2_block_group_alloc(osb, alloc_inode, bh);
450 if (status < 0) {
451 if (status != -ENOSPC)
452 mlog_errno(status);
453 goto bail;
455 atomic_inc(&osb->alloc_stats.bg_extends);
457 /* You should never ask for this much metadata */
458 BUG_ON(bits_wanted >
459 (le32_to_cpu(fe->id1.bitmap1.i_total)
460 - le32_to_cpu(fe->id1.bitmap1.i_used)));
463 get_bh(bh);
464 ac->ac_bh = bh;
465 bail:
466 if (bh)
467 brelse(bh);
469 mlog_exit(status);
470 return status;
473 int ocfs2_reserve_new_metadata(struct ocfs2_super *osb,
474 struct ocfs2_dinode *fe,
475 struct ocfs2_alloc_context **ac)
477 int status;
478 u32 slot;
480 *ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
481 if (!(*ac)) {
482 status = -ENOMEM;
483 mlog_errno(status);
484 goto bail;
487 (*ac)->ac_bits_wanted = ocfs2_extend_meta_needed(fe);
488 (*ac)->ac_which = OCFS2_AC_USE_META;
489 slot = osb->slot_num;
490 (*ac)->ac_group_search = ocfs2_block_group_search;
492 status = ocfs2_reserve_suballoc_bits(osb, (*ac),
493 EXTENT_ALLOC_SYSTEM_INODE, slot);
494 if (status < 0) {
495 if (status != -ENOSPC)
496 mlog_errno(status);
497 goto bail;
500 status = 0;
501 bail:
502 if ((status < 0) && *ac) {
503 ocfs2_free_alloc_context(*ac);
504 *ac = NULL;
507 mlog_exit(status);
508 return status;
511 int ocfs2_reserve_new_inode(struct ocfs2_super *osb,
512 struct ocfs2_alloc_context **ac)
514 int status;
516 *ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
517 if (!(*ac)) {
518 status = -ENOMEM;
519 mlog_errno(status);
520 goto bail;
523 (*ac)->ac_bits_wanted = 1;
524 (*ac)->ac_which = OCFS2_AC_USE_INODE;
526 (*ac)->ac_group_search = ocfs2_block_group_search;
528 status = ocfs2_reserve_suballoc_bits(osb, *ac,
529 INODE_ALLOC_SYSTEM_INODE,
530 osb->slot_num);
531 if (status < 0) {
532 if (status != -ENOSPC)
533 mlog_errno(status);
534 goto bail;
537 status = 0;
538 bail:
539 if ((status < 0) && *ac) {
540 ocfs2_free_alloc_context(*ac);
541 *ac = NULL;
544 mlog_exit(status);
545 return status;
548 /* local alloc code has to do the same thing, so rather than do this
549 * twice.. */
550 int ocfs2_reserve_cluster_bitmap_bits(struct ocfs2_super *osb,
551 struct ocfs2_alloc_context *ac)
553 int status;
555 ac->ac_which = OCFS2_AC_USE_MAIN;
556 ac->ac_group_search = ocfs2_cluster_group_search;
558 status = ocfs2_reserve_suballoc_bits(osb, ac,
559 GLOBAL_BITMAP_SYSTEM_INODE,
560 OCFS2_INVALID_SLOT);
561 if (status < 0 && status != -ENOSPC) {
562 mlog_errno(status);
563 goto bail;
566 bail:
567 return status;
570 /* Callers don't need to care which bitmap (local alloc or main) to
571 * use so we figure it out for them, but unfortunately this clutters
572 * things a bit. */
573 int ocfs2_reserve_clusters(struct ocfs2_super *osb,
574 u32 bits_wanted,
575 struct ocfs2_alloc_context **ac)
577 int status;
579 mlog_entry_void();
581 *ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
582 if (!(*ac)) {
583 status = -ENOMEM;
584 mlog_errno(status);
585 goto bail;
588 (*ac)->ac_bits_wanted = bits_wanted;
590 status = -ENOSPC;
591 if (ocfs2_alloc_should_use_local(osb, bits_wanted)) {
592 status = ocfs2_reserve_local_alloc_bits(osb,
593 bits_wanted,
594 *ac);
595 if ((status < 0) && (status != -ENOSPC)) {
596 mlog_errno(status);
597 goto bail;
598 } else if (status == -ENOSPC) {
599 /* reserve_local_bits will return enospc with
600 * the local alloc inode still locked, so we
601 * can change this safely here. */
602 mlog(0, "Disabling local alloc\n");
603 /* We set to OCFS2_LA_DISABLED so that umount
604 * can clean up what's left of the local
605 * allocation */
606 osb->local_alloc_state = OCFS2_LA_DISABLED;
610 if (status == -ENOSPC) {
611 status = ocfs2_reserve_cluster_bitmap_bits(osb, *ac);
612 if (status < 0) {
613 if (status != -ENOSPC)
614 mlog_errno(status);
615 goto bail;
619 status = 0;
620 bail:
621 if ((status < 0) && *ac) {
622 ocfs2_free_alloc_context(*ac);
623 *ac = NULL;
626 mlog_exit(status);
627 return status;
631 * More or less lifted from ext3. I'll leave their description below:
633 * "For ext3 allocations, we must not reuse any blocks which are
634 * allocated in the bitmap buffer's "last committed data" copy. This
635 * prevents deletes from freeing up the page for reuse until we have
636 * committed the delete transaction.
638 * If we didn't do this, then deleting something and reallocating it as
639 * data would allow the old block to be overwritten before the
640 * transaction committed (because we force data to disk before commit).
641 * This would lead to corruption if we crashed between overwriting the
642 * data and committing the delete.
644 * @@@ We may want to make this allocation behaviour conditional on
645 * data-writes at some point, and disable it for metadata allocations or
646 * sync-data inodes."
648 * Note: OCFS2 already does this differently for metadata vs data
649 * allocations, as those bitmaps are separate and undo access is never
650 * called on a metadata group descriptor.
652 static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh,
653 int nr)
655 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
657 if (ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap))
658 return 0;
659 if (!buffer_jbd(bg_bh) || !bh2jh(bg_bh)->b_committed_data)
660 return 1;
662 bg = (struct ocfs2_group_desc *) bh2jh(bg_bh)->b_committed_data;
663 return !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
666 static int ocfs2_block_group_find_clear_bits(struct ocfs2_super *osb,
667 struct buffer_head *bg_bh,
668 unsigned int bits_wanted,
669 unsigned int total_bits,
670 u16 *bit_off,
671 u16 *bits_found)
673 void *bitmap;
674 u16 best_offset, best_size;
675 int offset, start, found, status = 0;
676 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
678 if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
679 OCFS2_RO_ON_INVALID_GROUP_DESC(osb->sb, bg);
680 return -EIO;
683 found = start = best_offset = best_size = 0;
684 bitmap = bg->bg_bitmap;
686 while((offset = ocfs2_find_next_zero_bit(bitmap, total_bits, start)) != -1) {
687 if (offset == total_bits)
688 break;
690 if (!ocfs2_test_bg_bit_allocatable(bg_bh, offset)) {
691 /* We found a zero, but we can't use it as it
692 * hasn't been put to disk yet! */
693 found = 0;
694 start = offset + 1;
695 } else if (offset == start) {
696 /* we found a zero */
697 found++;
698 /* move start to the next bit to test */
699 start++;
700 } else {
701 /* got a zero after some ones */
702 found = 1;
703 start = offset + 1;
705 if (found > best_size) {
706 best_size = found;
707 best_offset = start - found;
709 /* we got everything we needed */
710 if (found == bits_wanted) {
711 /* mlog(0, "Found it all!\n"); */
712 break;
716 /* XXX: I think the first clause is equivalent to the second
717 * - jlbec */
718 if (found == bits_wanted) {
719 *bit_off = start - found;
720 *bits_found = found;
721 } else if (best_size) {
722 *bit_off = best_offset;
723 *bits_found = best_size;
724 } else {
725 status = -ENOSPC;
726 /* No error log here -- see the comment above
727 * ocfs2_test_bg_bit_allocatable */
730 return status;
733 static inline int ocfs2_block_group_set_bits(handle_t *handle,
734 struct inode *alloc_inode,
735 struct ocfs2_group_desc *bg,
736 struct buffer_head *group_bh,
737 unsigned int bit_off,
738 unsigned int num_bits)
740 int status;
741 void *bitmap = bg->bg_bitmap;
742 int journal_type = OCFS2_JOURNAL_ACCESS_WRITE;
744 mlog_entry_void();
746 if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
747 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
748 status = -EIO;
749 goto bail;
751 BUG_ON(le16_to_cpu(bg->bg_free_bits_count) < num_bits);
753 mlog(0, "block_group_set_bits: off = %u, num = %u\n", bit_off,
754 num_bits);
756 if (ocfs2_is_cluster_bitmap(alloc_inode))
757 journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
759 status = ocfs2_journal_access(handle,
760 alloc_inode,
761 group_bh,
762 journal_type);
763 if (status < 0) {
764 mlog_errno(status);
765 goto bail;
768 le16_add_cpu(&bg->bg_free_bits_count, -num_bits);
770 while(num_bits--)
771 ocfs2_set_bit(bit_off++, bitmap);
773 status = ocfs2_journal_dirty(handle,
774 group_bh);
775 if (status < 0) {
776 mlog_errno(status);
777 goto bail;
780 bail:
781 mlog_exit(status);
782 return status;
785 /* find the one with the most empty bits */
786 static inline u16 ocfs2_find_victim_chain(struct ocfs2_chain_list *cl)
788 u16 curr, best;
790 BUG_ON(!cl->cl_next_free_rec);
792 best = curr = 0;
793 while (curr < le16_to_cpu(cl->cl_next_free_rec)) {
794 if (le32_to_cpu(cl->cl_recs[curr].c_free) >
795 le32_to_cpu(cl->cl_recs[best].c_free))
796 best = curr;
797 curr++;
800 BUG_ON(best >= le16_to_cpu(cl->cl_next_free_rec));
801 return best;
804 static int ocfs2_relink_block_group(handle_t *handle,
805 struct inode *alloc_inode,
806 struct buffer_head *fe_bh,
807 struct buffer_head *bg_bh,
808 struct buffer_head *prev_bg_bh,
809 u16 chain)
811 int status;
812 /* there is a really tiny chance the journal calls could fail,
813 * but we wouldn't want inconsistent blocks in *any* case. */
814 u64 fe_ptr, bg_ptr, prev_bg_ptr;
815 struct ocfs2_dinode *fe = (struct ocfs2_dinode *) fe_bh->b_data;
816 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
817 struct ocfs2_group_desc *prev_bg = (struct ocfs2_group_desc *) prev_bg_bh->b_data;
819 if (!OCFS2_IS_VALID_DINODE(fe)) {
820 OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
821 status = -EIO;
822 goto out;
824 if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
825 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
826 status = -EIO;
827 goto out;
829 if (!OCFS2_IS_VALID_GROUP_DESC(prev_bg)) {
830 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, prev_bg);
831 status = -EIO;
832 goto out;
835 mlog(0, "Suballoc %llu, chain %u, move group %llu to top, prev = %llu\n",
836 (unsigned long long)le64_to_cpu(fe->i_blkno), chain,
837 (unsigned long long)le64_to_cpu(bg->bg_blkno),
838 (unsigned long long)le64_to_cpu(prev_bg->bg_blkno));
840 fe_ptr = le64_to_cpu(fe->id2.i_chain.cl_recs[chain].c_blkno);
841 bg_ptr = le64_to_cpu(bg->bg_next_group);
842 prev_bg_ptr = le64_to_cpu(prev_bg->bg_next_group);
844 status = ocfs2_journal_access(handle, alloc_inode, prev_bg_bh,
845 OCFS2_JOURNAL_ACCESS_WRITE);
846 if (status < 0) {
847 mlog_errno(status);
848 goto out_rollback;
851 prev_bg->bg_next_group = bg->bg_next_group;
853 status = ocfs2_journal_dirty(handle, prev_bg_bh);
854 if (status < 0) {
855 mlog_errno(status);
856 goto out_rollback;
859 status = ocfs2_journal_access(handle, alloc_inode, bg_bh,
860 OCFS2_JOURNAL_ACCESS_WRITE);
861 if (status < 0) {
862 mlog_errno(status);
863 goto out_rollback;
866 bg->bg_next_group = fe->id2.i_chain.cl_recs[chain].c_blkno;
868 status = ocfs2_journal_dirty(handle, bg_bh);
869 if (status < 0) {
870 mlog_errno(status);
871 goto out_rollback;
874 status = ocfs2_journal_access(handle, alloc_inode, fe_bh,
875 OCFS2_JOURNAL_ACCESS_WRITE);
876 if (status < 0) {
877 mlog_errno(status);
878 goto out_rollback;
881 fe->id2.i_chain.cl_recs[chain].c_blkno = bg->bg_blkno;
883 status = ocfs2_journal_dirty(handle, fe_bh);
884 if (status < 0) {
885 mlog_errno(status);
886 goto out_rollback;
889 status = 0;
890 out_rollback:
891 if (status < 0) {
892 fe->id2.i_chain.cl_recs[chain].c_blkno = cpu_to_le64(fe_ptr);
893 bg->bg_next_group = cpu_to_le64(bg_ptr);
894 prev_bg->bg_next_group = cpu_to_le64(prev_bg_ptr);
896 out:
897 mlog_exit(status);
898 return status;
901 static inline int ocfs2_block_group_reasonably_empty(struct ocfs2_group_desc *bg,
902 u32 wanted)
904 return le16_to_cpu(bg->bg_free_bits_count) > wanted;
907 /* return 0 on success, -ENOSPC to keep searching and any other < 0
908 * value on error. */
909 static int ocfs2_cluster_group_search(struct inode *inode,
910 struct buffer_head *group_bh,
911 u32 bits_wanted, u32 min_bits,
912 u16 *bit_off, u16 *bits_found)
914 int search = -ENOSPC;
915 int ret;
916 struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *) group_bh->b_data;
917 u16 tmp_off, tmp_found;
918 unsigned int max_bits, gd_cluster_off;
920 BUG_ON(!ocfs2_is_cluster_bitmap(inode));
922 if (gd->bg_free_bits_count) {
923 max_bits = le16_to_cpu(gd->bg_bits);
925 /* Tail groups in cluster bitmaps which aren't cpg
926 * aligned are prone to partial extention by a failed
927 * fs resize. If the file system resize never got to
928 * update the dinode cluster count, then we don't want
929 * to trust any clusters past it, regardless of what
930 * the group descriptor says. */
931 gd_cluster_off = ocfs2_blocks_to_clusters(inode->i_sb,
932 le64_to_cpu(gd->bg_blkno));
933 if ((gd_cluster_off + max_bits) >
934 OCFS2_I(inode)->ip_clusters) {
935 max_bits = OCFS2_I(inode)->ip_clusters - gd_cluster_off;
936 mlog(0, "Desc %llu, bg_bits %u, clusters %u, use %u\n",
937 (unsigned long long)le64_to_cpu(gd->bg_blkno),
938 le16_to_cpu(gd->bg_bits),
939 OCFS2_I(inode)->ip_clusters, max_bits);
942 ret = ocfs2_block_group_find_clear_bits(OCFS2_SB(inode->i_sb),
943 group_bh, bits_wanted,
944 max_bits,
945 &tmp_off, &tmp_found);
946 if (ret)
947 return ret;
949 /* ocfs2_block_group_find_clear_bits() might
950 * return success, but we still want to return
951 * -ENOSPC unless it found the minimum number
952 * of bits. */
953 if (min_bits <= tmp_found) {
954 *bit_off = tmp_off;
955 *bits_found = tmp_found;
956 search = 0; /* success */
960 return search;
963 static int ocfs2_block_group_search(struct inode *inode,
964 struct buffer_head *group_bh,
965 u32 bits_wanted, u32 min_bits,
966 u16 *bit_off, u16 *bits_found)
968 int ret = -ENOSPC;
969 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) group_bh->b_data;
971 BUG_ON(min_bits != 1);
972 BUG_ON(ocfs2_is_cluster_bitmap(inode));
974 if (bg->bg_free_bits_count)
975 ret = ocfs2_block_group_find_clear_bits(OCFS2_SB(inode->i_sb),
976 group_bh, bits_wanted,
977 le16_to_cpu(bg->bg_bits),
978 bit_off, bits_found);
980 return ret;
983 static int ocfs2_alloc_dinode_update_counts(struct inode *inode,
984 handle_t *handle,
985 struct buffer_head *di_bh,
986 u32 num_bits,
987 u16 chain)
989 int ret;
990 u32 tmp_used;
991 struct ocfs2_dinode *di = (struct ocfs2_dinode *) di_bh->b_data;
992 struct ocfs2_chain_list *cl = (struct ocfs2_chain_list *) &di->id2.i_chain;
994 ret = ocfs2_journal_access(handle, inode, di_bh,
995 OCFS2_JOURNAL_ACCESS_WRITE);
996 if (ret < 0) {
997 mlog_errno(ret);
998 goto out;
1001 tmp_used = le32_to_cpu(di->id1.bitmap1.i_used);
1002 di->id1.bitmap1.i_used = cpu_to_le32(num_bits + tmp_used);
1003 le32_add_cpu(&cl->cl_recs[chain].c_free, -num_bits);
1005 ret = ocfs2_journal_dirty(handle, di_bh);
1006 if (ret < 0)
1007 mlog_errno(ret);
1009 out:
1010 return ret;
1013 static int ocfs2_search_one_group(struct ocfs2_alloc_context *ac,
1014 handle_t *handle,
1015 u32 bits_wanted,
1016 u32 min_bits,
1017 u16 *bit_off,
1018 unsigned int *num_bits,
1019 u64 gd_blkno,
1020 u16 *bits_left)
1022 int ret;
1023 u16 found;
1024 struct buffer_head *group_bh = NULL;
1025 struct ocfs2_group_desc *gd;
1026 struct inode *alloc_inode = ac->ac_inode;
1028 ret = ocfs2_read_block(OCFS2_SB(alloc_inode->i_sb), gd_blkno,
1029 &group_bh, OCFS2_BH_CACHED, alloc_inode);
1030 if (ret < 0) {
1031 mlog_errno(ret);
1032 return ret;
1035 gd = (struct ocfs2_group_desc *) group_bh->b_data;
1036 if (!OCFS2_IS_VALID_GROUP_DESC(gd)) {
1037 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, gd);
1038 ret = -EIO;
1039 goto out;
1042 ret = ac->ac_group_search(alloc_inode, group_bh, bits_wanted, min_bits,
1043 bit_off, &found);
1044 if (ret < 0) {
1045 if (ret != -ENOSPC)
1046 mlog_errno(ret);
1047 goto out;
1050 *num_bits = found;
1052 ret = ocfs2_alloc_dinode_update_counts(alloc_inode, handle, ac->ac_bh,
1053 *num_bits,
1054 le16_to_cpu(gd->bg_chain));
1055 if (ret < 0) {
1056 mlog_errno(ret);
1057 goto out;
1060 ret = ocfs2_block_group_set_bits(handle, alloc_inode, gd, group_bh,
1061 *bit_off, *num_bits);
1062 if (ret < 0)
1063 mlog_errno(ret);
1065 *bits_left = le16_to_cpu(gd->bg_free_bits_count);
1067 out:
1068 brelse(group_bh);
1070 return ret;
1073 static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
1074 handle_t *handle,
1075 u32 bits_wanted,
1076 u32 min_bits,
1077 u16 *bit_off,
1078 unsigned int *num_bits,
1079 u64 *bg_blkno,
1080 u16 *bits_left)
1082 int status;
1083 u16 chain, tmp_bits;
1084 u32 tmp_used;
1085 u64 next_group;
1086 struct inode *alloc_inode = ac->ac_inode;
1087 struct buffer_head *group_bh = NULL;
1088 struct buffer_head *prev_group_bh = NULL;
1089 struct ocfs2_dinode *fe = (struct ocfs2_dinode *) ac->ac_bh->b_data;
1090 struct ocfs2_chain_list *cl = (struct ocfs2_chain_list *) &fe->id2.i_chain;
1091 struct ocfs2_group_desc *bg;
1093 chain = ac->ac_chain;
1094 mlog(0, "trying to alloc %u bits from chain %u, inode %llu\n",
1095 bits_wanted, chain,
1096 (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno);
1098 status = ocfs2_read_block(OCFS2_SB(alloc_inode->i_sb),
1099 le64_to_cpu(cl->cl_recs[chain].c_blkno),
1100 &group_bh, OCFS2_BH_CACHED, alloc_inode);
1101 if (status < 0) {
1102 mlog_errno(status);
1103 goto bail;
1105 bg = (struct ocfs2_group_desc *) group_bh->b_data;
1106 status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, bg);
1107 if (status) {
1108 mlog_errno(status);
1109 goto bail;
1112 status = -ENOSPC;
1113 /* for now, the chain search is a bit simplistic. We just use
1114 * the 1st group with any empty bits. */
1115 while ((status = ac->ac_group_search(alloc_inode, group_bh,
1116 bits_wanted, min_bits, bit_off,
1117 &tmp_bits)) == -ENOSPC) {
1118 if (!bg->bg_next_group)
1119 break;
1121 if (prev_group_bh) {
1122 brelse(prev_group_bh);
1123 prev_group_bh = NULL;
1125 next_group = le64_to_cpu(bg->bg_next_group);
1126 prev_group_bh = group_bh;
1127 group_bh = NULL;
1128 status = ocfs2_read_block(OCFS2_SB(alloc_inode->i_sb),
1129 next_group, &group_bh,
1130 OCFS2_BH_CACHED, alloc_inode);
1131 if (status < 0) {
1132 mlog_errno(status);
1133 goto bail;
1135 bg = (struct ocfs2_group_desc *) group_bh->b_data;
1136 status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, bg);
1137 if (status) {
1138 mlog_errno(status);
1139 goto bail;
1142 if (status < 0) {
1143 if (status != -ENOSPC)
1144 mlog_errno(status);
1145 goto bail;
1148 mlog(0, "alloc succeeds: we give %u bits from block group %llu\n",
1149 tmp_bits, (unsigned long long)le64_to_cpu(bg->bg_blkno));
1151 *num_bits = tmp_bits;
1153 BUG_ON(*num_bits == 0);
1156 * Keep track of previous block descriptor read. When
1157 * we find a target, if we have read more than X
1158 * number of descriptors, and the target is reasonably
1159 * empty, relink him to top of his chain.
1161 * We've read 0 extra blocks and only send one more to
1162 * the transaction, yet the next guy to search has a
1163 * much easier time.
1165 * Do this *after* figuring out how many bits we're taking out
1166 * of our target group.
1168 if (ac->ac_allow_chain_relink &&
1169 (prev_group_bh) &&
1170 (ocfs2_block_group_reasonably_empty(bg, *num_bits))) {
1171 status = ocfs2_relink_block_group(handle, alloc_inode,
1172 ac->ac_bh, group_bh,
1173 prev_group_bh, chain);
1174 if (status < 0) {
1175 mlog_errno(status);
1176 goto bail;
1180 /* Ok, claim our bits now: set the info on dinode, chainlist
1181 * and then the group */
1182 status = ocfs2_journal_access(handle,
1183 alloc_inode,
1184 ac->ac_bh,
1185 OCFS2_JOURNAL_ACCESS_WRITE);
1186 if (status < 0) {
1187 mlog_errno(status);
1188 goto bail;
1191 tmp_used = le32_to_cpu(fe->id1.bitmap1.i_used);
1192 fe->id1.bitmap1.i_used = cpu_to_le32(*num_bits + tmp_used);
1193 le32_add_cpu(&cl->cl_recs[chain].c_free, -(*num_bits));
1195 status = ocfs2_journal_dirty(handle,
1196 ac->ac_bh);
1197 if (status < 0) {
1198 mlog_errno(status);
1199 goto bail;
1202 status = ocfs2_block_group_set_bits(handle,
1203 alloc_inode,
1205 group_bh,
1206 *bit_off,
1207 *num_bits);
1208 if (status < 0) {
1209 mlog_errno(status);
1210 goto bail;
1213 mlog(0, "Allocated %u bits from suballocator %llu\n", *num_bits,
1214 (unsigned long long)le64_to_cpu(fe->i_blkno));
1216 *bg_blkno = le64_to_cpu(bg->bg_blkno);
1217 *bits_left = le16_to_cpu(bg->bg_free_bits_count);
1218 bail:
1219 if (group_bh)
1220 brelse(group_bh);
1221 if (prev_group_bh)
1222 brelse(prev_group_bh);
1224 mlog_exit(status);
1225 return status;
1228 /* will give out up to bits_wanted contiguous bits. */
1229 static int ocfs2_claim_suballoc_bits(struct ocfs2_super *osb,
1230 struct ocfs2_alloc_context *ac,
1231 handle_t *handle,
1232 u32 bits_wanted,
1233 u32 min_bits,
1234 u16 *bit_off,
1235 unsigned int *num_bits,
1236 u64 *bg_blkno)
1238 int status;
1239 u16 victim, i;
1240 u16 bits_left = 0;
1241 u64 hint_blkno = ac->ac_last_group;
1242 struct ocfs2_chain_list *cl;
1243 struct ocfs2_dinode *fe;
1245 mlog_entry_void();
1247 BUG_ON(ac->ac_bits_given >= ac->ac_bits_wanted);
1248 BUG_ON(bits_wanted > (ac->ac_bits_wanted - ac->ac_bits_given));
1249 BUG_ON(!ac->ac_bh);
1251 fe = (struct ocfs2_dinode *) ac->ac_bh->b_data;
1252 if (!OCFS2_IS_VALID_DINODE(fe)) {
1253 OCFS2_RO_ON_INVALID_DINODE(osb->sb, fe);
1254 status = -EIO;
1255 goto bail;
1257 if (le32_to_cpu(fe->id1.bitmap1.i_used) >=
1258 le32_to_cpu(fe->id1.bitmap1.i_total)) {
1259 ocfs2_error(osb->sb, "Chain allocator dinode %llu has %u used "
1260 "bits but only %u total.",
1261 (unsigned long long)le64_to_cpu(fe->i_blkno),
1262 le32_to_cpu(fe->id1.bitmap1.i_used),
1263 le32_to_cpu(fe->id1.bitmap1.i_total));
1264 status = -EIO;
1265 goto bail;
1268 if (hint_blkno) {
1269 /* Attempt to short-circuit the usual search mechanism
1270 * by jumping straight to the most recently used
1271 * allocation group. This helps us mantain some
1272 * contiguousness across allocations. */
1273 status = ocfs2_search_one_group(ac, handle, bits_wanted,
1274 min_bits, bit_off, num_bits,
1275 hint_blkno, &bits_left);
1276 if (!status) {
1277 /* Be careful to update *bg_blkno here as the
1278 * caller is expecting it to be filled in, and
1279 * ocfs2_search_one_group() won't do that for
1280 * us. */
1281 *bg_blkno = hint_blkno;
1282 goto set_hint;
1284 if (status < 0 && status != -ENOSPC) {
1285 mlog_errno(status);
1286 goto bail;
1290 cl = (struct ocfs2_chain_list *) &fe->id2.i_chain;
1292 victim = ocfs2_find_victim_chain(cl);
1293 ac->ac_chain = victim;
1294 ac->ac_allow_chain_relink = 1;
1296 status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits, bit_off,
1297 num_bits, bg_blkno, &bits_left);
1298 if (!status)
1299 goto set_hint;
1300 if (status < 0 && status != -ENOSPC) {
1301 mlog_errno(status);
1302 goto bail;
1305 mlog(0, "Search of victim chain %u came up with nothing, "
1306 "trying all chains now.\n", victim);
1308 /* If we didn't pick a good victim, then just default to
1309 * searching each chain in order. Don't allow chain relinking
1310 * because we only calculate enough journal credits for one
1311 * relink per alloc. */
1312 ac->ac_allow_chain_relink = 0;
1313 for (i = 0; i < le16_to_cpu(cl->cl_next_free_rec); i ++) {
1314 if (i == victim)
1315 continue;
1316 if (!cl->cl_recs[i].c_free)
1317 continue;
1319 ac->ac_chain = i;
1320 status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits,
1321 bit_off, num_bits, bg_blkno,
1322 &bits_left);
1323 if (!status)
1324 break;
1325 if (status < 0 && status != -ENOSPC) {
1326 mlog_errno(status);
1327 goto bail;
1331 set_hint:
1332 if (status != -ENOSPC) {
1333 /* If the next search of this group is not likely to
1334 * yield a suitable extent, then we reset the last
1335 * group hint so as to not waste a disk read */
1336 if (bits_left < min_bits)
1337 ac->ac_last_group = 0;
1338 else
1339 ac->ac_last_group = *bg_blkno;
1342 bail:
1343 mlog_exit(status);
1344 return status;
1347 int ocfs2_claim_metadata(struct ocfs2_super *osb,
1348 handle_t *handle,
1349 struct ocfs2_alloc_context *ac,
1350 u32 bits_wanted,
1351 u16 *suballoc_bit_start,
1352 unsigned int *num_bits,
1353 u64 *blkno_start)
1355 int status;
1356 u64 bg_blkno;
1358 BUG_ON(!ac);
1359 BUG_ON(ac->ac_bits_wanted < (ac->ac_bits_given + bits_wanted));
1360 BUG_ON(ac->ac_which != OCFS2_AC_USE_META);
1362 status = ocfs2_claim_suballoc_bits(osb,
1364 handle,
1365 bits_wanted,
1367 suballoc_bit_start,
1368 num_bits,
1369 &bg_blkno);
1370 if (status < 0) {
1371 mlog_errno(status);
1372 goto bail;
1374 atomic_inc(&osb->alloc_stats.bg_allocs);
1376 *blkno_start = bg_blkno + (u64) *suballoc_bit_start;
1377 ac->ac_bits_given += (*num_bits);
1378 status = 0;
1379 bail:
1380 mlog_exit(status);
1381 return status;
1384 int ocfs2_claim_new_inode(struct ocfs2_super *osb,
1385 handle_t *handle,
1386 struct ocfs2_alloc_context *ac,
1387 u16 *suballoc_bit,
1388 u64 *fe_blkno)
1390 int status;
1391 unsigned int num_bits;
1392 u64 bg_blkno;
1394 mlog_entry_void();
1396 BUG_ON(!ac);
1397 BUG_ON(ac->ac_bits_given != 0);
1398 BUG_ON(ac->ac_bits_wanted != 1);
1399 BUG_ON(ac->ac_which != OCFS2_AC_USE_INODE);
1401 status = ocfs2_claim_suballoc_bits(osb,
1403 handle,
1406 suballoc_bit,
1407 &num_bits,
1408 &bg_blkno);
1409 if (status < 0) {
1410 mlog_errno(status);
1411 goto bail;
1413 atomic_inc(&osb->alloc_stats.bg_allocs);
1415 BUG_ON(num_bits != 1);
1417 *fe_blkno = bg_blkno + (u64) (*suballoc_bit);
1418 ac->ac_bits_given++;
1419 status = 0;
1420 bail:
1421 mlog_exit(status);
1422 return status;
1425 /* translate a group desc. blkno and it's bitmap offset into
1426 * disk cluster offset. */
1427 static inline u32 ocfs2_desc_bitmap_to_cluster_off(struct inode *inode,
1428 u64 bg_blkno,
1429 u16 bg_bit_off)
1431 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1432 u32 cluster = 0;
1434 BUG_ON(!ocfs2_is_cluster_bitmap(inode));
1436 if (bg_blkno != osb->first_cluster_group_blkno)
1437 cluster = ocfs2_blocks_to_clusters(inode->i_sb, bg_blkno);
1438 cluster += (u32) bg_bit_off;
1439 return cluster;
1442 /* given a cluster offset, calculate which block group it belongs to
1443 * and return that block offset. */
1444 u64 ocfs2_which_cluster_group(struct inode *inode, u32 cluster)
1446 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1447 u32 group_no;
1449 BUG_ON(!ocfs2_is_cluster_bitmap(inode));
1451 group_no = cluster / osb->bitmap_cpg;
1452 if (!group_no)
1453 return osb->first_cluster_group_blkno;
1454 return ocfs2_clusters_to_blocks(inode->i_sb,
1455 group_no * osb->bitmap_cpg);
1458 /* given the block number of a cluster start, calculate which cluster
1459 * group and descriptor bitmap offset that corresponds to. */
1460 static inline void ocfs2_block_to_cluster_group(struct inode *inode,
1461 u64 data_blkno,
1462 u64 *bg_blkno,
1463 u16 *bg_bit_off)
1465 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1466 u32 data_cluster = ocfs2_blocks_to_clusters(osb->sb, data_blkno);
1468 BUG_ON(!ocfs2_is_cluster_bitmap(inode));
1470 *bg_blkno = ocfs2_which_cluster_group(inode,
1471 data_cluster);
1473 if (*bg_blkno == osb->first_cluster_group_blkno)
1474 *bg_bit_off = (u16) data_cluster;
1475 else
1476 *bg_bit_off = (u16) ocfs2_blocks_to_clusters(osb->sb,
1477 data_blkno - *bg_blkno);
1481 * min_bits - minimum contiguous chunk from this total allocation we
1482 * can handle. set to what we asked for originally for a full
1483 * contig. allocation, set to '1' to indicate we can deal with extents
1484 * of any size.
1486 int __ocfs2_claim_clusters(struct ocfs2_super *osb,
1487 handle_t *handle,
1488 struct ocfs2_alloc_context *ac,
1489 u32 min_clusters,
1490 u32 max_clusters,
1491 u32 *cluster_start,
1492 u32 *num_clusters)
1494 int status;
1495 unsigned int bits_wanted = max_clusters;
1496 u64 bg_blkno = 0;
1497 u16 bg_bit_off;
1499 mlog_entry_void();
1501 BUG_ON(ac->ac_bits_given >= ac->ac_bits_wanted);
1503 BUG_ON(ac->ac_which != OCFS2_AC_USE_LOCAL
1504 && ac->ac_which != OCFS2_AC_USE_MAIN);
1506 if (ac->ac_which == OCFS2_AC_USE_LOCAL) {
1507 status = ocfs2_claim_local_alloc_bits(osb,
1508 handle,
1510 bits_wanted,
1511 cluster_start,
1512 num_clusters);
1513 if (!status)
1514 atomic_inc(&osb->alloc_stats.local_data);
1515 } else {
1516 if (min_clusters > (osb->bitmap_cpg - 1)) {
1517 /* The only paths asking for contiguousness
1518 * should know about this already. */
1519 mlog(ML_ERROR, "minimum allocation requested %u exceeds "
1520 "group bitmap size %u!\n", min_clusters,
1521 osb->bitmap_cpg);
1522 status = -ENOSPC;
1523 goto bail;
1525 /* clamp the current request down to a realistic size. */
1526 if (bits_wanted > (osb->bitmap_cpg - 1))
1527 bits_wanted = osb->bitmap_cpg - 1;
1529 status = ocfs2_claim_suballoc_bits(osb,
1531 handle,
1532 bits_wanted,
1533 min_clusters,
1534 &bg_bit_off,
1535 num_clusters,
1536 &bg_blkno);
1537 if (!status) {
1538 *cluster_start =
1539 ocfs2_desc_bitmap_to_cluster_off(ac->ac_inode,
1540 bg_blkno,
1541 bg_bit_off);
1542 atomic_inc(&osb->alloc_stats.bitmap_data);
1545 if (status < 0) {
1546 if (status != -ENOSPC)
1547 mlog_errno(status);
1548 goto bail;
1551 ac->ac_bits_given += *num_clusters;
1553 bail:
1554 mlog_exit(status);
1555 return status;
1558 int ocfs2_claim_clusters(struct ocfs2_super *osb,
1559 handle_t *handle,
1560 struct ocfs2_alloc_context *ac,
1561 u32 min_clusters,
1562 u32 *cluster_start,
1563 u32 *num_clusters)
1565 unsigned int bits_wanted = ac->ac_bits_wanted - ac->ac_bits_given;
1567 return __ocfs2_claim_clusters(osb, handle, ac, min_clusters,
1568 bits_wanted, cluster_start, num_clusters);
1571 static inline int ocfs2_block_group_clear_bits(handle_t *handle,
1572 struct inode *alloc_inode,
1573 struct ocfs2_group_desc *bg,
1574 struct buffer_head *group_bh,
1575 unsigned int bit_off,
1576 unsigned int num_bits)
1578 int status;
1579 unsigned int tmp;
1580 int journal_type = OCFS2_JOURNAL_ACCESS_WRITE;
1581 struct ocfs2_group_desc *undo_bg = NULL;
1583 mlog_entry_void();
1585 if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
1586 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
1587 status = -EIO;
1588 goto bail;
1591 mlog(0, "off = %u, num = %u\n", bit_off, num_bits);
1593 if (ocfs2_is_cluster_bitmap(alloc_inode))
1594 journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
1596 status = ocfs2_journal_access(handle, alloc_inode, group_bh,
1597 journal_type);
1598 if (status < 0) {
1599 mlog_errno(status);
1600 goto bail;
1603 if (ocfs2_is_cluster_bitmap(alloc_inode))
1604 undo_bg = (struct ocfs2_group_desc *) bh2jh(group_bh)->b_committed_data;
1606 tmp = num_bits;
1607 while(tmp--) {
1608 ocfs2_clear_bit((bit_off + tmp),
1609 (unsigned long *) bg->bg_bitmap);
1610 if (ocfs2_is_cluster_bitmap(alloc_inode))
1611 ocfs2_set_bit(bit_off + tmp,
1612 (unsigned long *) undo_bg->bg_bitmap);
1614 le16_add_cpu(&bg->bg_free_bits_count, num_bits);
1616 status = ocfs2_journal_dirty(handle, group_bh);
1617 if (status < 0)
1618 mlog_errno(status);
1619 bail:
1620 return status;
1624 * expects the suballoc inode to already be locked.
1626 int ocfs2_free_suballoc_bits(handle_t *handle,
1627 struct inode *alloc_inode,
1628 struct buffer_head *alloc_bh,
1629 unsigned int start_bit,
1630 u64 bg_blkno,
1631 unsigned int count)
1633 int status = 0;
1634 u32 tmp_used;
1635 struct ocfs2_super *osb = OCFS2_SB(alloc_inode->i_sb);
1636 struct ocfs2_dinode *fe = (struct ocfs2_dinode *) alloc_bh->b_data;
1637 struct ocfs2_chain_list *cl = &fe->id2.i_chain;
1638 struct buffer_head *group_bh = NULL;
1639 struct ocfs2_group_desc *group;
1641 mlog_entry_void();
1643 if (!OCFS2_IS_VALID_DINODE(fe)) {
1644 OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
1645 status = -EIO;
1646 goto bail;
1648 BUG_ON((count + start_bit) > ocfs2_bits_per_group(cl));
1650 mlog(0, "%llu: freeing %u bits from group %llu, starting at %u\n",
1651 (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno, count,
1652 (unsigned long long)bg_blkno, start_bit);
1654 status = ocfs2_read_block(osb, bg_blkno, &group_bh, OCFS2_BH_CACHED,
1655 alloc_inode);
1656 if (status < 0) {
1657 mlog_errno(status);
1658 goto bail;
1661 group = (struct ocfs2_group_desc *) group_bh->b_data;
1662 status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, group);
1663 if (status) {
1664 mlog_errno(status);
1665 goto bail;
1667 BUG_ON((count + start_bit) > le16_to_cpu(group->bg_bits));
1669 status = ocfs2_block_group_clear_bits(handle, alloc_inode,
1670 group, group_bh,
1671 start_bit, count);
1672 if (status < 0) {
1673 mlog_errno(status);
1674 goto bail;
1677 status = ocfs2_journal_access(handle, alloc_inode, alloc_bh,
1678 OCFS2_JOURNAL_ACCESS_WRITE);
1679 if (status < 0) {
1680 mlog_errno(status);
1681 goto bail;
1684 le32_add_cpu(&cl->cl_recs[le16_to_cpu(group->bg_chain)].c_free,
1685 count);
1686 tmp_used = le32_to_cpu(fe->id1.bitmap1.i_used);
1687 fe->id1.bitmap1.i_used = cpu_to_le32(tmp_used - count);
1689 status = ocfs2_journal_dirty(handle, alloc_bh);
1690 if (status < 0) {
1691 mlog_errno(status);
1692 goto bail;
1695 bail:
1696 if (group_bh)
1697 brelse(group_bh);
1699 mlog_exit(status);
1700 return status;
1703 int ocfs2_free_dinode(handle_t *handle,
1704 struct inode *inode_alloc_inode,
1705 struct buffer_head *inode_alloc_bh,
1706 struct ocfs2_dinode *di)
1708 u64 blk = le64_to_cpu(di->i_blkno);
1709 u16 bit = le16_to_cpu(di->i_suballoc_bit);
1710 u64 bg_blkno = ocfs2_which_suballoc_group(blk, bit);
1712 return ocfs2_free_suballoc_bits(handle, inode_alloc_inode,
1713 inode_alloc_bh, bit, bg_blkno, 1);
1716 int ocfs2_free_clusters(handle_t *handle,
1717 struct inode *bitmap_inode,
1718 struct buffer_head *bitmap_bh,
1719 u64 start_blk,
1720 unsigned int num_clusters)
1722 int status;
1723 u16 bg_start_bit;
1724 u64 bg_blkno;
1725 struct ocfs2_dinode *fe;
1727 /* You can't ever have a contiguous set of clusters
1728 * bigger than a block group bitmap so we never have to worry
1729 * about looping on them. */
1731 mlog_entry_void();
1733 /* This is expensive. We can safely remove once this stuff has
1734 * gotten tested really well. */
1735 BUG_ON(start_blk != ocfs2_clusters_to_blocks(bitmap_inode->i_sb, ocfs2_blocks_to_clusters(bitmap_inode->i_sb, start_blk)));
1737 fe = (struct ocfs2_dinode *) bitmap_bh->b_data;
1739 ocfs2_block_to_cluster_group(bitmap_inode, start_blk, &bg_blkno,
1740 &bg_start_bit);
1742 mlog(0, "want to free %u clusters starting at block %llu\n",
1743 num_clusters, (unsigned long long)start_blk);
1744 mlog(0, "bg_blkno = %llu, bg_start_bit = %u\n",
1745 (unsigned long long)bg_blkno, bg_start_bit);
1747 status = ocfs2_free_suballoc_bits(handle, bitmap_inode, bitmap_bh,
1748 bg_start_bit, bg_blkno,
1749 num_clusters);
1750 if (status < 0)
1751 mlog_errno(status);
1753 mlog_exit(status);
1754 return status;
1757 static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg)
1759 printk("Block Group:\n");
1760 printk("bg_signature: %s\n", bg->bg_signature);
1761 printk("bg_size: %u\n", bg->bg_size);
1762 printk("bg_bits: %u\n", bg->bg_bits);
1763 printk("bg_free_bits_count: %u\n", bg->bg_free_bits_count);
1764 printk("bg_chain: %u\n", bg->bg_chain);
1765 printk("bg_generation: %u\n", le32_to_cpu(bg->bg_generation));
1766 printk("bg_next_group: %llu\n",
1767 (unsigned long long)bg->bg_next_group);
1768 printk("bg_parent_dinode: %llu\n",
1769 (unsigned long long)bg->bg_parent_dinode);
1770 printk("bg_blkno: %llu\n",
1771 (unsigned long long)bg->bg_blkno);
1774 static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe)
1776 int i;
1778 printk("Suballoc Inode %llu:\n", (unsigned long long)fe->i_blkno);
1779 printk("i_signature: %s\n", fe->i_signature);
1780 printk("i_size: %llu\n",
1781 (unsigned long long)fe->i_size);
1782 printk("i_clusters: %u\n", fe->i_clusters);
1783 printk("i_generation: %u\n",
1784 le32_to_cpu(fe->i_generation));
1785 printk("id1.bitmap1.i_used: %u\n",
1786 le32_to_cpu(fe->id1.bitmap1.i_used));
1787 printk("id1.bitmap1.i_total: %u\n",
1788 le32_to_cpu(fe->id1.bitmap1.i_total));
1789 printk("id2.i_chain.cl_cpg: %u\n", fe->id2.i_chain.cl_cpg);
1790 printk("id2.i_chain.cl_bpc: %u\n", fe->id2.i_chain.cl_bpc);
1791 printk("id2.i_chain.cl_count: %u\n", fe->id2.i_chain.cl_count);
1792 printk("id2.i_chain.cl_next_free_rec: %u\n",
1793 fe->id2.i_chain.cl_next_free_rec);
1794 for(i = 0; i < fe->id2.i_chain.cl_next_free_rec; i++) {
1795 printk("fe->id2.i_chain.cl_recs[%d].c_free: %u\n", i,
1796 fe->id2.i_chain.cl_recs[i].c_free);
1797 printk("fe->id2.i_chain.cl_recs[%d].c_total: %u\n", i,
1798 fe->id2.i_chain.cl_recs[i].c_total);
1799 printk("fe->id2.i_chain.cl_recs[%d].c_blkno: %llu\n", i,
1800 (unsigned long long)fe->id2.i_chain.cl_recs[i].c_blkno);